Instagram | Sponsor this newsletter | Let’s Connect!

My fellow AI explorers

I’m not one to scream "the sky is falling"—but after watching Eric Schmidt calmly explain that computers no longer need to listen to us… I had to pause.

This week’s edition is heavy. Not with doom, but with clarity. A former Google CEO, a groundbreaking new model from OpenAI, and a surprise top-tier image generator from ByteDance are painting a future that’s arriving faster than we expected.

And while everyone’s distracted by TikTok bans and ChatGPT hacks, something far more transformative is happening right under our noses.

In today’s edition:

• 🧠 The Rise of Superintelligence — And Why Eric Schmidt Thinks It’s Underhyped

• 🧪 OpenAI’s O3 Model Might Be the Smartest LLM on Earth

• 🎨 ByteDance’s Seedream 3: Quietly Beating GPT-4o at Image Generation

Must See AI Tools

• 💰 Payman: AI That Pays Humans. Over 10,000+ signed up for the beta

• 💫 SubMagic: An AI tool that edits short-form content for you! (Get 10% off using code “uncoverai” at checkout)

• 🎤 11Labs: #1 AI voice generator (Click Here to get 10,000 free credits upon signing up!)

• 🤖 ManyChat: Automate your responses & conversations on IG, FB and more! (Click Here to get first month for free)

• 🎙️ Syllaby: The only social media marketing tool you’ll ever need - powered by AI! (Get 25% off the first month or any annual plan with code “UNCOVER” at checkout)

Get discounts on popular AI tools here!

Eric Schmidt

🧠 The Future of AI, According to Eric Schmidt

Artificial Superintelligence is coming—and we’re not ready.

In a gripping panel, Eric Schmidt (ex-Google CEO) dropped a quiet bomb: superintelligent systems may arrive in as little as 6 years. That’s not a sci-fi pitch—it’s the current San Francisco consensus, driven by pure scaling.

Here’s what stood out:

• Self-improving agents are already writing 10–20% of their own code in frontier labs.

• AI coders will write 90–100% of code within 12 months, according to OpenAI insiders.

• AI mathematicians are incoming — ones smarter than top grad students.

This isn’t hype. Schmidt calls ASI underhyped because the implications are so staggering that they’re hard to internalize. If AI hits this level, society as we know it will change—permanently.

He didn’t stop there:

• Biotech meets AI: Schmidt is backing startups where foundation models suggest drug hypotheses by night—and robotic labs test them by morning.

• Longevity leap: The real goal? Map every druggable human target within 2 years—putting us on the brink of life extension breakthroughs.

• Agentic takeover: Schmidt imagines buying a house with AI agents handling every step—from finding the land to suing the contractor.

💡 Takeaway: Recursive self-improvement isn't theory anymore—it's in motion. And if these systems start designing better versions of themselves, we’re heading into a world where humans may no longer understand what our machines are doing... but still use them every day.

▶️ Watch the full Eric Schmidt conversation

AI Images

🎨 Seedream 3: ByteDance’s Silent Image Powerhouse

It came out of nowhere—and it's beating GPT-4o in realism and style.

ByteDance just launched Seedream 3, and the image results are nothing short of jaw-dropping. On Artificial Analysis’ leaderboard, it’s tied with GPT-4o for the top spot—and in some use-cases, it looks even better.

Why this matters:

• Better realism: Slightly blurry, imperfect details make images look more real, less synthetic.

• Human anatomy accuracy: Consistently nails tricky poses—like bent limbs, foreshortening, or group dynamics.

• Less censorship: Unlike OpenAI’s image model, it can generate likenesses of real people (think Will Smith, Queen Elizabeth) with surprising accuracy.

What impressed us:

• Anime, Pixar, and Monet-style renderings are smooth, stylistic, and consistent.

• 3D isometric scenes (like a full bedroom layout) looked more natural than GPT-4o’s polished but rigid output.

• Real-world tests: From fake low-res selfies to retro student parties, Seedream often felt more photographic than AI-made.

⚠️ Caveat: Seedream struggles with long text and image editing (think comic panels or receipts). It’s not a poster design tool—yet.

🖼️ Verdict: If you want realism or raw creative flair, Seedream 3 may be the most fun and capable image generator available—especially since it’s free to use with 150 daily credits.

🎨 Try Seedream at dreamina.capcut.com

Open AI

🧪 O3: OpenAI’s New Genius-Level Model

O3 isn’t just smart—it might be the smartest model on the planet.

OpenAI’s latest language model, internally referred to as “O3,” recently aced the Mensa IQ test with a score of 136. That’s well above Gemini 2.5 Pro’s 128 and far ahead of GPT-4’s predecessors.

Here’s why O3 is different:

• Tool use within its own chain of thought — it writes, runs, and interprets code in real time.

• Zero-shot reasoning: It solved complex math benchmarks faster than any human solver.

• Geoguessing God Mode: It pinpointed a rural Quebec road from a random Street View image… in seconds.

O3 doesn’t just mimic experts. It becomes one—crafting novel scientific theories, explaining post-labor economics, and even solving mazes and visual logic puzzles flawlessly.

📈 Benchmarks:

• #1 spot on the Artificial Analysis Intelligence Index

• Best-in-class for math, coding, and reasoning across nearly every test

• Almost no hallucinations, even in high-pressure, high-context tasks

🧠 Prediction: This model will power the next generation of autonomous agents—models that don’t just respond but think, plan, and act with expert-level precision.

See how O3 performs in this full breakdown

30-Second AI Play

🎧 30-Second AI Play: Clone Yourself (in Art Style)

Want to remix your selfie into anime, Pixar, or Manet? Here’s how using Seedream:

1. Go to dreamina.capcut.com and generate or upload a base image.

2. Click “Reference” → Select “Character” or “Human Face”

3. Choose your art style in the prompt (e.g., “Pixar 3D portrait of this face”)

4. Hit Generate — and watch the transformation happen

🧙 Why this is cool:

• Works with your own photos

• Styles like Studio Ghibli, Monet, and Anime feel hand-crafted

• Faster and more flexible than Midjourney or DALL·E for most style applications

💡 Pro tip: You can use it to create a character brand for yourself—or produce stylized avatars across different social platforms in seconds.

Try it out here

Other Relevant AI News!

💸 OpenAI is in talks to acquire Windsurf for ~$3B — a rival AI coding assistant platform, in what could be one of its largest acquisitions yet.
🧠 GPT-4.1 is here — OpenAI’s new flagship model brings faster performance, longer memory, and better reasoning (all within ChatGPT).
🤝 Before Windsurf, OpenAI explored a deal with Cursor — showing how serious it is about dominating the coding agent space.
🖥️ Microsoft's new Recall feature for Windows previews lets users access “memories” of everything they’ve done on their PC — powered by local AI.
🗣️ Anthropic is building a voice assistant to compete with OpenAI — pushing Claude into the race for smart, always-on AI companions.

Golden Nuggets

• ⚙️ Superintelligence isn’t coming—it’s forming, quietly, in recursive loops

• 🧠 O3 is pushing AI from tool… to collaborator

• 🎨 ByteDance might secretly own the best image model on the planet

Until our next AI rendezvous,

Anthony | Founder of Uncover AI